[nskelly]

Hi,

I was wondering what methods are there to initialize a DNS flow field? I am looking at the flow within a cone and plate rheometer and I am unsure of what ICs or OF utilities are best for adding some sore turbulent intensity or perturbations. Currently, the only methods I have seen are perturbU, but this is restricted to channel flows.

Thanks

[Andrea1984]

Hi Nathaniel,

I have obtained an initial isotropic turbulent condition for LES with boxTurb. As the name suggests, the tool only works for box-shaped domains, and there is the additional caveat that the mesh has to have 2^n cells. For non-box shaped domains with a generic number of cells I usually do the following:

-generate a box with 2^n cells which encompasses the domain
-use boxTurb to generate a nice turbulent field within the box
-use mapFields with the cuttingPatches option to map the turbulent field to my actual domain

Hope this helps,
Andrea

[nskelly]

Hi Andrea,

Thanks for the reply. I read that using boxTurb is one method for initialising the flow. But I am a bit unsure of how the mapFields part works.

I want to look at the laminar-turbulent transition process. So how can I add the turbulent fluctuations to an existing simulation? The

For example:
If I want to run a simulation at Re = 2000 (where flow is in the transitional flow regime). Where do I put the the boxTurb field data? After some timesteps have solved or as the inital BCs?

Thanks

[Andrea1984]

I have never used boxTurb to superimpose turbulent fluctuations to an existing velocity field. The way I would use it is to generate an initial turbulent flow field and then run the simulation and let the "real" turbulence develop.

I am not sure this is the right thing to do in your case though, since you want to investigate transition and boxTurb is going to provide you with a fully turbulent initial flow field.

Andrea